The Geochemistry and Bioturbation of Clay Sediments Associated with Amalgamated Crusts at the Gagua Ridge

Based on the analysis of geochemical and mineralogical compositions, deep sea clay sediment characteristics and their material sources were examined in the eastern flank of the Gagua Ridge. The mineralogy mainly consists of detrital clay minerals, quartz, and authigenic phillipsite. There is scarce biogenic debris (siliceous or calcareous). The consolidated sediments are more enriched in Si, Al, K, Na, Li, Sc, Cr, Rb, and Cs than the associated crusts and nodules. The unmixed sediment samples were mainlycontributed by Asian eolian dust. The onset of the outer Fe-Mn crust growth nearly coincides with the Central Asia aridification event at ~3.5 Ma, which resulted in an abrupt increase in eolian flux of Asian dust. Intensified surface primary productivity is assumed to bring more metals to deep waters, and eventually facilitate the outer Fe-Mn crust formation. Authigenic phillipsite may come from the alteration of local basic volcanic glasses and cause excess Al, high Al/Ti, and low Si/Al ratios. However, phillipsites hardly affect the abundance of rare earth elements (REEs) and their patterns. In addition, the investigation of two kinds of burrows inside the consolidated sediments reveals that the inner nodules of the amalgamated crusts may remain on the oxic sediment surface, due to frequent benthic activities.


Introduction
Fe-Mn nodules form by precipitation from both seawater and sediment pore waters on the sediment surface.Fe-Mn oxide crusts grow on rock surfaces throughout the global oceans, at water depths ranging from approximately 400-7000 m [1].Crusts also grow on the hardground of consolidated or indurated sediments, such as the amalgamated crusts from the Shatsky Rise and Gagua Ridge [2,3], and the Fe-Mn crusts in the Central Indian Basin [4,5].In addition, the amalgamated crusts also show a close proximity to the underlying nodules cemented by the consolidated sediments.Deep sea Fe-Mn crusts, nodules, and their associated sediments have been investigated by many researchers studying interelemental relationships and hoping to understand the possible contributions from different sources (i.e., [6][7][8][9][10][11][12]). In the Pacific Ocean, Elderfield et al. (1981) [6] observed that the highest nodule rare earth elements (REEs) are associated with the lowest REEs of the associated sediments, and the nodules with the largest positive Ce anomalies are found on sediments with the smallest negative Ce anomalies, indicating competitive scavenging of REEs between the nodules and sediments.However, such a relation is not found within Central Indian Ocean Basin nodule-sediment pairs [11] or in the southwest Pacific Ocean [7].Pattan and Parthiban (2011) [11] concluded that it may not be appropriate to correlate elemental behavior between nodule and sediment pairs, because none of the major, trace, and rare earth elements exhibit any interelemental relationships between these pairs.
Major and trace elements in nodules, micronodules, and abyssal clay on the central Clarion-Clipperton abyssal plain presented complex signatures, reflecting the influence of both the upper continental crust and MORB, as well as probable East Pacific Rise material transport via the Pacific North Equatorial Current [10].Banakar et al. (2003) [13] also identified a Himalayan-derived silicate-detritus component in a Central Indian Ocean Fe-Mn crust that was located nearly 1000 km south of the Bengal Fan.Therefore, the Fe-Mn crusts are also expected to have recorded the past changes in sediment supply from the Ganges and Brahmaputra River systems to the Bengal Fan.
The north-south oriented Gagua Ridge is a continuous bathymetric high ridge that extends over 300 km from the Island of Luzon in the Philippines to the north and intersects the Ryukyu subduction zone east of Taiwan [14,15].The Gagua Ridge separates the Western Philippine Basin (WPB) to the east from the Huatung Basin (HB) to the west.Amalgamated Fe-Mn crusts were first recovered in 2016 at the eastern flank of the Gagua Ridge, and these crusts have been analyzed for their geochemical and mineralogical compositions in order to understand the formation process [3].The amalgamated crust was composed of three layers: an outer crust, a middle crust, and inner nodules.All samples were mainly composed of Fe-rich vernadite, amorphous FeOOH and associated quartz, plagioclase, pyroxene and Mg-titanomagnetite.Consolidated sediments filled in the space between the outer crust and inner nodules, and the middle crust grew into the sediments [3].There are also many burrows in the sediment samples with different sizes and fillings.Previous studies have reported the occurrence of ubiquitous benthic faunal activities on crust and nodule-covered deep sea sediments, such as tracks, trails, and burrowing structures [4,5,[16][17][18][19][20][21].The intense and prolonged bioturbation activity of the benthic fauna may have facilitated the occurrence of Fe-Mn nodules and micronodules [5,20].
In the present study, major, trace and, rare earth elemental analyses and mineralogical examinations were conducted for clay sediments covered by an outer Fe-Mn crust.The objectives of this study were to understand the possible sediment provenance, their significance to elemental behavior in sediments and associated crusts and nodules, and to explore the role of bioturbation in the formation of the amalgamated crusts in the Gagua Ridge.

Background
The study area is located on the eastern flank of the Gagua Ridge, which lies on the westernmost edge of the WPB (Figure 1).The Gagua Ridge is affected by two water masses off the coast of southeastern Taiwan (Figure 1).One water mass is the northward-flowing Kuroshio Current (KC), and the other water mass is the southward-flowing Luzon Undercurrent (LUC).At 22 • N-25 • N, the KC is approximately 300 m deep and 170 km wide, with a maximum velocity of 1 m/s and a volume transport between 15 and 25 Sv [22].Therefore, the transport of sediments delivered by rivers from mainland China and the Taiwan mountain belt would not easily reach the Gagua Ridge.Although thick sedimentary units (~4 km) exist in the HB [23], these units have become ponded within the basin, due to the presence of the Gagua Ridge.The VM33-95 core collected in the HB was estimated to have a sedimentation rate of 6 cm/kyr [24].The sedimentation rate to the east of the Gagua Ridge in the WPB is low (approximately 5.3-5.4mm/kyr) according to the nearest stations, PH-6 and St 6, from Huh (1992) [25].Asian continental dust may contribute approximately 10-50% of the detrital fraction of sediments on the Benham Rise, which lies ~300 km south of the Gagua Ridge in the WPB [26].Thus, the main source of material in the Gagua Ridge would be a combination of eolian dust and volcanic debris [3].
Minerals 2019, 9, x FOR PEER REVIEW 3 of 16 [26].Thus, the main source of material in the Gagua Ridge would be a combination of eolian dust and volcanic debris [3].

Materials and Methods
The amalgamated Fe-Mn crusts were collected at a water depth of 4071 m by TV-grab during the HOBAB4 cruise in 2016.The indurated sediments in this study existed between inner nodules and outer crusts (Figure 2a).The sediment samples were fine-grained clayey sediments that possessed many burrowing structures (Figure 2a,b).Fine grains of Fe-Mn oxides aggregate inside some large burrows and were also dispersed in the sediments.

Materials and Methods
The amalgamated Fe-Mn crusts were collected at a water depth of 4071 m by TV-grab during the HOBAB4 cruise in 2016.The indurated sediments in this study existed between inner nodules and outer crusts (Figure 2a).The sediment samples were fine-grained clayey sediments that possessed many burrowing structures (Figure 2a,b).Fine grains of Fe-Mn oxides aggregate inside some large burrows and were also dispersed in the sediments.Sediments beneath the outer Fe-Mn crust were cut, washed with distilled water, oven-dried at 60 °C and ground in an agate mortar.Bulk powder samples were investigated by X-ray diffraction, using a Bruker X-ray diffractometer (XRD) with a Cu-target at the Institute of Oceanology, Chinese Academy of Sciences (Qingdao, China).The results were analyzed using the MDI Jade software (Version 6, Materials Data, Inc., Livermore, CA, USA) to measure X-ray reflections and distinguish possible mineral compositions.The major element compositions were determined by X-ray fluorescence (XRF) at the ALS laboratory, Guangzhou.The XRF analysis was performed in Sediments beneath the outer Fe-Mn crust were cut, washed with distilled water, oven-dried at 60 • C and ground in an agate mortar.Bulk powder samples were investigated by X-ray diffraction, using a Bruker X-ray diffractometer (XRD) with a Cu-target at the Institute of Oceanology, Chinese Academy of Sciences (Qingdao, China).The results were analyzed using the MDI Jade software (Version 6, Materials Data, Inc., Livermore, CA, USA) to measure X-ray reflections and distinguish possible mineral compositions.The major element compositions were determined by X-ray fluorescence (XRF) at the ALS laboratory, Guangzhou.The XRF analysis was performed in conjunction with loss-on-ignition at 1000 • C. The accuracy of the analytical procedure was assessed using certified reference materials (SARM-5; NCSDC47009; and GBW07295).The precision of the analysis was greater than ±5%.Trace element and rare earth element concentrations were determined using a PE ELAN DRC II inductively coupled plasma mass spectrometer (ICP-MS) at the Institute of Oceanology, Chinese Academy of Sciences.The accuracy based on standard reference materials (GBW07315 and GBW07316) was greater than ±5%, and the precision based on a duplicate sample analysis was ±5%.
The bulk sediments and burrows were also observed using a Tescan scanning electron microscope (SEM) after being coated with gold at the Institute of Oceanology, Chinese Academy of Sciences (Qingdao, China).The SEM system used combines a TESCAN VEGA 3 LMH SEM with an Oxford INCA X-Max energy dispersive spectrometer (EDS).This system has an energy resolution of 124 eV (Mn Kα), is capable of collecting count rates at >500,000 cps with a throughput >200,000 cps and is equipped with both a YAG-based backscatter electron (BSE) detector and a secondary electron detector.Some minerals were determined by EDS attached to the SEM.

Mineralogy
The sediments came from below the carbonate compensation depth (CCD), and none of the calcareous debris remained.There were hardly any biogenic siliceous remnants, such as radiolarians, based on SEM observations.The sediments mainly consisted of clay minerals, quartz, and phillipsite, with minor amounts of plagioclase, ilmenite, augite, zircon, and apatite (Figure 2).The clay minerals were very fine-grained and predominantly composed of illite, as identified on the XRD diagram (Figure 3).The crystal fragments of volcanogenic minerals were the largest, and reached up to 300 µm in diameter, such as in the cases of plagioclase and ilmenite (Figure 2).Trace zircon and carbonate fluorapatite were observed by SEM (see Supplementary Figure S1).The size of the prismatic authigenic phillipsite fell in the middle.Phillipsite is common in sediments, but rarely appears in associated crusts and nodules [3].The phillipsite peaks usually overlap with those of kaolinite/chlorite at 7.2 Å, but the 7.2 Å peak here was largely attributed to phillipsite, as indicated by the very weak intensity of the 3.5 Å peak for kaolinite/chlorite compared to the strong intensity of the 7.2 Å peak.
Minerals 2019, 9, x FOR PEER REVIEW 5 of 16 conjunction with loss-on-ignition at 1000 °C.The accuracy of the analytical procedure was assessed using certified reference materials (SARM-5; NCSDC47009; and GBW07295).The precision of the analysis was greater than ±5%.Trace element and rare earth element concentrations were determined using a PE ELAN DRC II inductively coupled plasma mass spectrometer (ICP-MS) at the Institute of Oceanology, Chinese Academy of Sciences.The accuracy based on standard reference materials (GBW07315 and GBW07316) was greater than ±5%, and the precision based on a duplicate sample analysis was ±5%.
The bulk sediments and burrows were also observed using a Tescan scanning electron microscope (SEM) after being coated with gold at the Institute of Oceanology, Chinese Academy of Sciences (Qingdao, China).The SEM system used combines a TESCAN VEGA 3 LMH SEM with an Oxford INCA X-Max energy dispersive spectrometer (EDS).This system has an energy resolution of 124 eV (Mn Kα), is capable of collecting count rates at >500,000 cps with a throughput >200,000 cps and is equipped with both a YAG-based backscatter electron (BSE) detector and a secondary electron detector.Some minerals were determined by EDS attached to the SEM.

Mineralogy
The sediments came from below the carbonate compensation depth (CCD), and none of the calcareous debris remained.There were hardly any biogenic siliceous remnants, such as radiolarians, based on SEM observations.The sediments mainly consisted of clay minerals, quartz, and phillipsite, with minor amounts of plagioclase, ilmenite, augite, zircon, and apatite (Figure 2).The clay minerals were very fine-grained and predominantly composed of illite, as identified on the XRD diagram (Figure 3).The crystal fragments of volcanogenic minerals were the largest, and reached up to 300 μm in diameter, such as in the cases of plagioclase and ilmenite (Figure 2).Trace zircon and carbonate fluorapatite were observed by SEM (see Supplementary Figure S1).The size of the prismatic authigenic phillipsite fell in the middle.Phillipsite is common in sediments, but rarely appears in associated crusts and nodules [3].The phillipsite peaks usually overlap with those of kaolinite/chlorite at 7.2 Å, but the 7.2 Å peak here was largely attributed to phillipsite, as indicated by the very weak intensity of the 3.5 Å peak for kaolinite/chlorite compared to the strong intensity of the 7.2 Å peak.

Bulk Chemical Compositions
In the Gagua Ridge samples, the mean contents of Si, Al, K, and Na in the sediments were higher than those of the associated crusts and nodules (see Table 1), and Ca and Mg were in a similar range in the sediments and crusts and nodules.The Fe, Mn, Ti, and P of the sediments were lower than those of the associated crusts and nodules.The crusts and nodules were more enriched in most of the trace elements than the sediments, whereas the Li, Sc, Cr, Rb and Cs contents were enriched in the sediments.In the sediments, manganese was largely present in the form of a scattered Fe-Mn oxide phase.Two sediment samples (R3-1e and R3-2e) were mixed with certain amounts of Fe-Mn oxides, which were shown by remarkably elevated contents of Co and Mn (mean 251 ppm and 1.97 wt.%, respectively) and a high Co/Th ratio (mean 10.55).The other six samples were considered to be unmixed, with trace amounts of oxides and average Co and Mn contents of 48 ppm and 0.48 wt.%, respectively.
The average Al/(Al + Fe + Mn) ratio of 0.66 was identical to that of the upper continental crust (UCC) value (0.67; [27]), while the Al/Ti ratio of 26.5 of the unmixed sediment samples from this area was higher than the UCC value (21.2; [27]).
The Si/Al ratio of all sediment samples was uniform, with an average of 2.42, which was close to that of the deep sea authigenic phillipsite (2.3-2.8;[28,29]), but smaller than that of the associated crusts and nodules (3.03), UCC (3.82), Luzon volcanics (2.71), Taiwan volcanics (3.0), south Ryukyu Islands (3.2), and Chinese loess (4.06) (See Table 1).The Co/Th ratio of unmixed samples with an average of 4.84 was higher than the values of UCC and loess, but lower than the value of volcanogenic materials from adjacent islands.The La/Sc and Th/Hf ratios, which can be used to identify the UCC debris contribution [30], were close to the values of the UCC and loess, but dramatically higher than those of the volcanogenic materials from adjacent islands.

REE Abundance and Pattern
The total REE abundances (∑REEs) of unmixed sediment samples were between 165 and 201 ppm, and the two oxide-mixed samples (R3-1e and R3-2e) were 254 and 406 ppm, respectively.The ∑REEs in the associated nodules and crusts (average of 1264 ppm) were much higher, due to the adsorption of REEs by Fe-Mn oxides [3].REE plus Y (REY) patterns of unmixed sediment samples normalized to post-Archean Australian shale (PAAS, [31]) show weak negative Ce anomalies (0.67-0.83), positive Eu anomalies (1.21-1.27),and small enrichments in middle REY, while REY patterns normalized to chondrite [32] show enrichments in light REY and negative anomalies for both Ce and Eu (Figure 4b).The two mixed samples showed an overall similar pattern as those of crusts and nodules, with positive anomalies of Ce and negative anomalies of Eu and Y.The mixing of Fe-Mn oxides in mixed sediments not only produced higher total amounts of REY, but also enabled the positive Ce anomalies.Eu and Ce anomalies discussed in this paper are depicted as (Eu/Eu*) PAAS = Eu PAAS /(Sm PAAS * Gd PAAS ) 1/2 and (Ce/Ce*) PAAS = Ce PAAS /(La PAAS * Pr PAAS ) 1/2 , respectively.The Pliocene red clay of the Chinese Loess Plateau [33], volcanic rocks from Mount Arayat of Luzon [34], Iriomote-jima Island in south Ryukyu [35], the Coastal Range of Taiwan [36,37], and North Pacific bottom seawater (~4000 m depth, [38]) are plotted for comparison (Figure 4), demonstrating the similarity between the unmixed sediments and the Jiaxian red clay in the Chinese Loess Plateau.

Excess Si and Al in the Sediments
Ti is generally considered to be continentally derived, and is the most stable element under varying physicochemical changes.Therefore, assuming that all Ti is mineralogically bound to the terrigenous fraction, Ti was used to calculate the terrigenous percentages in sediments and the excess of multiple elements (e.g., Al, Fe, Mn, and Ba), in order to assess the relative contributions of terrigenous and nonterrigenous components [11,[39][40][41].However, Ti may be enriched in hydrogenetic crusts and nodules due to effective surface complexation from the water column on oxide surfaces [42], suggesting that this method may not work in the two mixed sediment samples, because the samples contain a remarkable amount of fine Fe-Mn oxides.As a result, we calculated the excess of Al for only the unmixed samples, using the equation Al ex = Al − ((Al PAAS /Ti PAAS ) × Ti) after Murray and Leinen (1996) [39], and calculated the excess of Si using Si ex = SiO 2 − 3.38 × Al 2 O 3 [11].

Oxidation-Reduction Conditions
The redox potential of the depositional environment can be determined using the parameter (Mn*) of the surface sediments after Machhour et al. (1994) [43]: Mn* = log[(Mn samples /Mn shales )/(Fe samples /Fe shales )], with a positive value of Mn* indicating an oxidizing environment and a negative value indicating a suboxic or reducing environment.The mean values for Mn shales (0.085 wt.%) and Fe shales (4.72 wt.%) were taken from Li and Schoonmaker (2014) [44].In the eastern flank of the Gagua Ridge, all the samples have positive Mn* values, with an average of 0.84.Alternatively, the V/Cr ratio can also be applied for determining the redox condition of the sediment [41,45], where values of >4.5 indicate an anoxic environment, and <2 indicates an oxic environment.All our sediment samples showed values of <2.

Bioturbated Burrows
In the Gagua Ridge, there is abundant evidence of bioturbation inside the consolidated sediments, including two kinds of burrows (a large burrow is at least 6-7 centimeters long and ~5 mm wide, and a small burrow is 2-3 centimeters long and ~2 mm wide).The large burrows are lined by black micrograins of vernadite, and the small burrows are filled with dark gray sediments.According to Knaust (2017) [46], the large burrows with Fe-Mn oxide linings may be Thalassinoides, and the small burrows with mud infillings may be grouped as Planolites.No trails of recent biota have been observed from the large burrow walls.In Figure 2, one large burrow crosses a small burrow that is also lined with oxides, indicating an earlier occurrence of small burrows.Additionally, small burrows may have been isolated from oxic seawater soon after formation, and thus avoided Fe-Mn oxide accretion.

Contributions from Possible Sources
The clay sediment samples were assumed to have accumulated earlier than 3.8 Ma, due to the empirical dating of the outer Fe-Mn crust [3].As the Philippine Sea Plate and the Gagua Ridge are subducted into the Eurasian Plate at a rate of ~71 km/Ma [47], the sediments have likely been forming in the southeast ~300 km away from the present location since the Pliocene.Reportedly, the northern Pacific sediments were mainly derived from the Taklimakan desert through long-distance transportation by a westerly wind [48,49].In contrast, surface clay sediments in the eastern Philippine Basin (e.g., [50]) were mainly derived from the alteration of local and nearby volcanic materials, and the detrital fraction of the sediments from the Benham Rise [26,51] comprised the Asian dust contribution, which varied between ~10% to ~50% via the East Asian winter monsoon (EAWM).We employ the ternary diagram of Sc-La-Th, which has been successfully used in eolian deposit provenance identifications (e.g., [52][53][54][55]), and discrimination between eolian and volcanogenic inputs to the Pacific sediments [56,57].We also adopt the Eu/Sm-Sc/Th plot to display the geochemical domains of the sediment sample sources (Figure 5).The unmixed sediments in the Gagua Ridge were used to interpret the possible sources, since the two mixed samples are evidently biased towards the distinct domain of Fe-Mn crusts and nodules.Both diagrams indicate the close proximity of our samples to Asian dust (which is represented by the Pliocene red clay in the Chinese Loess Plateau), demonstrating the main input of eolian dust via the EAWM.This eolian dust input is also evidenced by the abundant illite in the clay minerals (Figure 3), and the REE abundances and patterns similar to that of Asian loess (Figure 4).
As deep sea ferromanganese deposits and sediments receive metals through different processes and accumulate on different time scales (a million years versus a millennium), Pattan and Parthiban (2011) [11] argued that it is not appropriate to correlate any elemental behavior between nodules and their associated sediments.However, the formation of amalgamated crusts from the Gagua Ridge may still be affected by certain sedimentation events (e.g., the sharp increase in eolian flux).The principal iron source to the surface ocean is eolian dust [58].The ecosystem may be significantly influenced by the iron supply from eolian dust in the tropical western Pacific [59].Xu et al. (2015) [51] suggested that the Asian dust input notably affected marine productivity in the western Philippine Sea, and the higher levels of dust input may be a key process for enhancing biological productivity.In the study area, the onset of the outer crust growth (~3.8Ma) at the eastern flank of the Gagua Ridge nearly coincides with the Central Asia aridification event at ~3.5 Ma, which was primarily caused by global cooling [60], and resulted in an abrupt increasing eolian flux of Asian dust [60][61][62].Consequently, it can be speculated that intensified surface primary productivity that occurred at that time and may produce more particulate organic matter that was exported to deep waters and carried more metals (such as Mn, Co, Ni and Cu) downward [63].This process would facilitate the formation of the outer Fe-Mn crust, due to the dissolution of biogenic materials at the Gagua Ridge.Meanwhile, the deposition of dust on the seafloor would fill in the pore spaces and cement the Fe-Mn oxides and volcanic detritus [3].
to the Pacific sediments [56,57].We also adopt the Eu/Sm-Sc/Th plot to display the geochemical domains of the sediment sample sources (Figure 5).The unmixed sediments in the Gagua Ridge were used to interpret the possible sources, since the two mixed samples are evidently biased towards the distinct domain of Fe-Mn crusts and nodules.Both diagrams indicate the close proximity of our samples to Asian dust (which is represented by the Pliocene red clay in the Chinese Loess Plateau), demonstrating the main input of eolian dust via the EAWM.This eolian dust input is also evidenced by the abundant illite in the clay minerals (Figure 3), and the REE abundances and patterns similar to that of Asian loess (Figure 4).
As deep sea ferromanganese deposits and sediments receive metals through different processes and accumulate on different time scales (a million years versus a millennium), Pattan and Parthiban (2011) [11] argued that it is not appropriate to correlate any elemental behavior between nodules and their associated sediments.However, the formation of amalgamated crusts from the Gagua Ridge may still be affected by certain sedimentation events (e.g., the sharp increase in eolian flux).The principal iron source to the surface ocean is eolian dust [58].The ecosystem may be significantly influenced by the iron supply from eolian dust in the tropical western Pacific [59].Xu et al. (2015) [51] suggested that the Asian dust input notably affected marine productivity in the western Philippine Sea, and the higher levels of dust input may be a key process for enhancing biological productivity.In the study area, the onset of the outer crust growth (~3.8Ma) at the eastern flank of the Gagua Ridge nearly coincides with the Central Asia aridification event at ~3.5 Ma, which was primarily caused by global cooling [60], and resulted in an abrupt increasing eolian flux of Asian dust [60][61][62].Consequently, it can be speculated that intensified surface primary productivity that occurred at that time and may produce more particulate organic matter that was exported to deep waters and carried more metals (such as Mn, Co, Ni and Cu) downward [63].This process would facilitate the formation of the outer Fe-Mn crust, due to the dissolution of biogenic materials at the Gagua Ridge.Meanwhile, the deposition of dust on the seafloor would fill in the pore spaces and cement the Fe-Mn oxides and volcanic detritus [3].[33], Bau and Knittel (1993) [34], Shinjo (1999) [35] and Lai et al. (2008;2017) [36,37], respectively.Sc contents, Th contents and Eu/Sm ratios of mid-North Pacific seawater from ~4000 m depth are from Amakawa et al. (2007) [64], Roy-Barman et al. (1996) [65], and Zhang and Nozaki (1996) [38], respectively.
The unmixed sediment samples have no Si ex , which is present as biogenic opal, and the Al ex in the sediment varies between 3.6% and 4.2%, while crusts and nodules do not contain Al ex .The Si ex in outer crusts (average 3.7%) was higher than that of the inner nodules (average 1.0%), while the middle crusts had hardly any Si ex , indicating that the sedimentation environment may have changed at the transition of the inner-middle and middle-outer interface of the amalgamated crust.The Fe-Mn layers of inner nodules and outer crusts may have retained a small amount of radiolarian shells.Due to the absence of any biogenic material (e.g., radiolarian shells or diatoms) in the sediments, the presence of the Al ex could not be due to scavenging of dissolved Al by the biogenic components [39,66], but may be contributed to by the presence of volcanogenic materials and authigenic phillipsite.
Volcanogenic materials from Luzon Island and eastern Taiwan may be transported to the eastern flank of the Gagua Ridge by the northward KC and south-flowing LUC.However, this process is carried out in a subordinate style, which is shown in the discrimination diagram (Figure 5).The data were gathered in the boundary of the loess domain and have no sign of binary mixing between the volcanogenic material from nearby islands and Asian dust.Little contribution from the south Ryukyu Islands to the study site was indicated.Large crystals of euhedral ilmenite and crystal fragments of pyroxene and plagioclase were found, with poor roundness, under SEM observations (Figure 2d,e), indicating the erosional input by local igneous basement rocks.The Eu anomaly may be derived from either volcanogenic material or from eolian dust in terms of different normalizing patterns (positive Eu anomalies in Figure 4a or negative Eu anomalies in Figure 4b).In addition, the authigenic phillipsite accounts for a considerable proportion of the sediments in the eastern flank of the Gagua Ridge (Figure 3), leading to high Al/Ti and low Si/Al ratios.Deep sea phillipsite may form by diagenesis, hydrothermalism, metamorphism, and halmyrolysis processes [29].The phillipsite in our samples is most likely related to halmyrolysis, which altered local basic volcanic glasses from the Gagua Ridge.
Phillipsite may play an important role in the REY enrichment of REY-rich muds in addition to Fe and Mn-oxyhydroxides [67].However, Dubinin and Sval'nov (2000) [68] noted that the phillipsite from the Southern Basin of the Pacific Ocean did not absorb REEs from deep ocean water.The REE composition of phillipsite, similar to that of shales, was possibly formed under the combined influence of the exchange complexes of solid phases (clay minerals and iron oxyhydroxides) in the sediment [69].The volcaniclastic REE signatures were not inherited by authigenic phillipsite during complete zeolitization in oxidized sediments [68].Therefore, the phillipsite aggregated with REY-containing concomitants (including biogenic apatite, Fe-oxyhydroxide, and Fe-Ca-hydroxophosphate) was responsible for the high REE values in REY-rich muds [70].Toyoda et al. (1990) [71] reported a large negative Ce anomaly for pelagic sediments in the equatorial Pacific, which was related to fish bone debris composed of biogenic calcium phosphate.The strong negative Ce anomaly in the red clay from the Wharton basin was also attributed to the presence of authigenic smectite, fish bone teeth and phillipsite [72].For our sediment samples, we inferred that the small negative Ce anomalies in unmixed sediment samples may have originated from trace amounts of biogenic apatite that accumulate REEs with a strong negative Ce anomaly.

Bottom Sediment Bioturbation and Its Implications
Two kinds of burrows inside the consolidated sediments of the Gagua Ridge indicate different faunal activities.The soft and indurated sediments may be continuously disturbed and reworked by bioturbation [5,20].Surface nodules can also be lifted and frequently turned over by benthic animals, as observed in the KODOS area [73], the equatorial North Pacific [18], the Central Indian Basin [5], and the Peru Basin [19].In the Gagua Ridge, the inner nodules were likely kept on the sediment surface due to frequent benthic activities before the middle Fe-Mn crust initiation.The nodules on the sediments can also be pushed downward by benthic activities, although an upward push may be easier [18].As proposed by Chen et al. (2018) [3], when the inner nodules sank into the sediments, the middle Fe-Mn crust of the amalgamated crusts may have grown hydrogenetically under oxic conditions from the surface of the inner nodules into the sediments.Positive Mn* values and V/Cr ratio values (<2) of the Gagua Ridge sediment samples, which are related to the redox conditions [41], indicate an oxic environment.The dissolved oxygen may be supplied from bottom water continuously into the sediment column through void burrows, which may produce the same solution environment as the bottom seawater and obstruct the diagenetic growth of 10-Å manganates.
The large Thalassinoides-like burrows with Mn oxides coating the burrows consisted of Ni-rich Mn oxides [17], todorokite [16], or vernadite in this study (Figure 2), which may provide material for the accretion of macro manganese nodules or crusts.Jung and Lee (1999) [20] considered that Mn oxide burrow linings likely formed by the reprecipitation of metal ions remobilized from surrounding oxygen-depleted sediments adjacent to the burrows.However, vernadite growing inside the burrows in the Gagua Ridge samples indicates direct hydrogenetic precipitation under oxic conditions.The small Planolites-like burrows with active mud in-filling [46,74] are the same as those occurring in the Central Indian Basin [4,5] and may form as a result of back-filling by benthic fauna.The bioturbation should gradually cease before the accretion of the outer crust, while Fe-Mn coatings inside the burrows may form in between.

Conclusions
We established the bulk geochemical and mineralogical data of the sediments associated with the amalgamated crusts of the eastern flank of the Gagua Ridge.The consolidated sediments were covered by an outer Fe-Mn crust, with an age of approximately 3.8 Ma.The major and trace elements of the sediments and associated crusts and nodules showed different enrichment behaviors.Two samples were mixed with certain amounts of scattered Fe-Mn oxide particles, and the other samples with negligible mixing were adopted for the provenance analysis.The REE pattern and bulk geochemical diagrams (e.g., Sc-La-Th and Eu/Sm-Sc/Th) displayed the possible sources of our samples.The Asian eolian dust was assumed to be the main source of the detrital input, with a minor contribution from local volcanogenic materials.The abrupt increasing eolian flux corresponding to the Central Asia aridification event at ~3.5 Ma may have facilitated the outer Fe-Mn crust formation.The abundant authigenic phillipsite may have been formed by the alteration of local basic volcanic glasses and resulted in the excess Al, high Al/Ti, and low Si/Al ratios.The negative Ce anomalies in the unmixed samples may have been caused by trace apatite minerals.The sediments were deposited under oxic conditions with abundant burrow structures.Large Thalassinoides-like burrows were coated with black hydrogenetic Fe-Mn oxide grains, and small Planolites-like burrows were backfilled by the benthic fauna.The benthic activities may have maintained the inner nodules on the sediment surface, and the burrows could provide a hydrogenetic environment for the accretion of vernadite pillars on the burrow wall and facilitate the growth of the middle crust into sediments.

Figure 2 .
Figure 2. (a,b) Hand specimens of consolidated sediment samples; (c,d) SEM images of sediments; (eh) SEM images of the interior of the burrows.(a) Indurated clay sediments underlie an outer Fe-Mn crust.Fine particles of Fe-Mn oxides are observed scattering inside the sediments.(b) Planolites-like and Thalassinoides-like burrow structures in a fragment of hard sediments.The Thalassinoides-like burrows are coated by black Fe-Mn oxides.Burrows are indicated by arrows.(c) The debris of detrital illite.(d) Large ilmenite (ilm) minerals trapped in the sediment matrix and prisms of phillipsite (phil) in the cavities.(e) Large fragment of plagioclase (plag) crystal (~300 μm long) trapped in the boundary between the sediments and oxide coatings.(f) Overview of the interior of the burrow structure.A fragment of plagioclase crystal drops among the oxide pillars.(g) The upward growth of a columnar structure of hydrogenetic Fe-Mn oxide indicated by the arrow.The spheroidal surface morphology is caused by dozens of microspheres (each diameter is ~60 μm).(h)The cross section of one microsphere accreting from the sediment matrix.

Figure 2 .
Figure 2. (a,b) Hand specimens of consolidated sediment samples; (c,d) SEM images of sediments; (e-h) SEM images of the interior of the burrows.(a) Indurated clay sediments underlie an outer Fe-Mn crust.Fine particles of Fe-Mn oxides are observed scattering inside the sediments.(b) Planolites-like and Thalassinoides-like burrow structures in a fragment of hard sediments.The Thalassinoides-like burrows are coated by black Fe-Mn oxides.Burrows are indicated by arrows.(c) The debris of detrital illite.(d) Large ilmenite (ilm) minerals trapped in the sediment matrix and prisms of phillipsite (phil) in the cavities.(e) Large fragment of plagioclase (plag) crystal (~300 µm long) trapped in the boundary between the sediments and oxide coatings.(f) Overview of the interior of the burrow structure.A fragment of plagioclase crystal drops among the oxide pillars.(g) The upward growth of a columnar structure of hydrogenetic Fe-Mn oxide indicated by the arrow.The spheroidal surface morphology is caused by dozens of microspheres (each diameter is ~60 µm).(h) The cross section of one microsphere accreting from the sediment matrix.

Figure 3 .
Figure 3. Representative XRD pattern of one bulk sediment sample from the Gagua Ridge.Quartz, plagioclase, phillipsite, and illite can be identified in the diagram.

Figure 3 .
Figure 3. Representative XRD pattern of one bulk sediment sample from the Gagua Ridge.Quartz, plagioclase, phillipsite, and illite can be identified in the diagram.

Table 1 .
Major element (wt.%) and trace element (mg/kg) data of sediment samples in the eastern flank of the Gagua Ridge.Data of associated crust & nodule samples are also listed.